x4 has come full circle. Announced in 2006, x4 made headlines as a technological breakthrough.
By 2007, x4 was generally viewed as a problematic disappointment.
Here we are in late 2008 and x4 is back in fashion, once again being trumpeted as a technological breakthrough. Curiously, the company putting x4 back in the headlines is none other than Samsung, SanDisk’s arch-rival, and ardent suitor.
Samsung has even put a value on x4- “billions of dollars.”
From SanDisk’s recent arbitration win over Samsung:
“According to the testimony at the hearings, an early termination of the license caused by a termination of the Strategic Agreement will have serious financial consequences. This is due to M-Systems announcement in the second quarter of 2006 of a technological breakthrough called x4 technology, which enables, for the first time, the use of 4 bits per cell NAND. This effectively doubles the capacity of the 2 bits per cell NAND technology now utilized in flash memory products. According to the testimony, this new technology has significant implications for personal computers and all entertainment devices. The timing of M-Systems Notice of Termination did not escape Samsung who argues strenuously that, without the license granted under the Strategic Agreement, it will only be able to acquire the rights to the new x4 technology through a renewal or extension of its existing technology cross-license with SanDisk. According to Samsung, SanDisk will use the leverage gained from a victory in this arbitration to force additional royalties that could potentially cost Samsung billions of dollars.”
So What’s So Great About x4?
There’s more to x4 than meets the eye. Most folks seem to think of x4 simply as a slightly bulked up version of 3 bit/cell MLC (x3.) If it were that simple, Samsung would never have uttered “billions.”
When msystems announced x4, some experts in the business came right out and said msystems claims of 4 bits/cell performance and reliability simply weren’t possible.
In fact, a slightly bulked up version of x3 probably isn’t workable. Too squirrelly.
msystems succeeded at the so-called impossible 4 bits/cell by coming at the problem from an entirely different angle. x4 isn’t a straightforward implementation of 4 bits/cell. Its another beast altogether. How it works has never been revealed, but I’ve got my guesses.
Before speculating further, some background and back-story are in order.
Background and Back-Story
As discussed, SanDisk comes by x4 via its acquisition of msystems in July 2006.
x4 was announced on msystems one and only analyst day in May 2006. This msystems analyst day was webcast live. Luckily I recorded the event. No replay was ever available
These presentations include x4 info that to my knowledge has never been presented elsewhere. The presentations were done in two parts, I’ve posted transcriptions of both. Part I has most of the good stuff. Particularly recommend Dov’s comments, Simon Litsyn’s presentation and the Q&A:
My post “What happened to x4?” attempts to update the x4 saga within the context of July 2007. In hindsight, this post still holds up, but misses the x4 alternative reality (joke).
My post “Stratosphere and x4”, in many respects, is the inspiration for this post. The most important thing to remember when reading this post is that my basic thesis is wrong. x4 is a subset of Stratosphere. The real deal is revealed in Amir Ban’s comments following the post. As he put it:
“Interesting and intelligent analysis, but the 7,023,735 patent is not the “x4 patent”. Indeed, a search for an “x4 patent” must come up empty-handed, since x4 is a straightforward application of Stratosphere. Stratosphere envisions taking some baseline flash and increasing bits/cell, using methods described in 6,469,931 and later patents. x4 means using the same methods when the baseline is 2 bit/cell MLC and the target is 4 bit/cell.”
Many thanks Amir.
x4 is, in many respects, a misnomer. The fundamental technology isn’t 4 bit/cell specific. The original name for the technology was Stratosphere. To my knowledge this name is no longer used. x4 came later. As Amir put it:
“The term x4 was coined by my Stratosphere group in a meeting with Toshiba in March 2004 to describe the specific cooperation proposal we were putting forward to them.”
While it is true that x4 is a technology that when applied to regular MLC (2 bits/cell) doubles storage capacity to 4 bits/cell, it can be also be applied to SLC resulting in x1.58±. or can be applied to x3 resulting in x5± (possibly).
It can also be used to improve NAND reliability, and isn’t necessarily limited to flash memory.
The key x4 technology patent appears to be U.S. patent #6,469,931, “Method for Increasing Information Content in a Computer Memory.” This patent was applied for in 2001 and issued in 2002. Inventors are listed as Amir Ban, Simon Litsyn and Idan Alrod. Assignee is msystems.
The key idea is that groups of flash memory cells can store information above and beyond the information stored in the flash memory cells themselves. With a few twists, a top-up in memory capacity is possible for “free.” The image below, from the patent, illustrates a schematic description of a two-dimensional optimal encoding scheme.
One example given in the patent is for single level cell flash memory. Today, 16 cells of SLC NAND can store 16 bits of information. Using x4 technology, 20 bits of information can be stored using the same 16 cells of SLC- a 25% gain in capacity.
Capacity gains increase when more memory cells are used, reaching a limit of around 58% for SLC. Hence x4 technology can theoretically turn x1 (SLC) into x1.58.
Capacity gains increase as the bits/cell increase as well. When x4 technology is applied to 2 bits/cell MLC, 4 bits of information can be stored. Hence the name x4.
Theoretically x4 can be applied to 3 bits/cell resulting in something close to x5, but this may not be practical as x3 already uses aggressive coding and the voltage levels may not be stable.
A second fundamental key x4 technology patent appears to be U.S. patent #7,023,735, “Method for Increasing the Reliability of a Flash Memory.” This patent was applied for in 2004 and issued in 2006. Inventors are listed as Amir Ban, Simon Litsyn and Idan Alrod. Assignee is Ramot at Tel-Aviv University.
The main idea is that by programming flash memory using reference voltages, fractional reference voltages, and redundancy bits, the reliability of info stored in MLC flash memory can be improved. Again, just like patent #6,469,931, each memory cell is not considered only by itself. Reliability can be improved by taking advantage of the frame of reference offered by groups of memory cells.
I can’t help but wonder if SanDisk hasn’t found a use for x4 technology beyond the the dedicated 4 bit/cell chip. Clearly the technology itself would seem to lend itself to many applications. The first which comes to mind is MLC SSDs- Not 4 bit/cell MLC, but rather 2 bit/cell SSDs with improved performance and reliability. That be sweet- if improvements were significant.
This might help explain Samsung’s contention that x4 will prove important enough to cost Samsung billions of dollars in licensing fees. If its something else that’s OK too. Inquiring minds want to know.
SanDisk’s lawyers may yet figure out a way to bring Samsung’s testimony public. Maybe then we’ll have a better idea of why Samsung thinks x4 is so great.